Advanced digital image forensics toolkit for analyzing evidence, detecting AI generation, and identifying image manipulations. Built with Streamlit and modern AI models (CLIP, ViT).
- 🖼️ Similarity Search: Find visually and semantically similar images in the database via image upload or text keywords (using CLIP).
- 🤖 AI Generation Detection: Identify if an image is AI-generated (GAN/Diffusion) with confidence scores.
- 🛠️ Manipulation Analysis: Perform Error Level Analysis (ELA) and statistical attribute analysis to detect tampering.
- 📄 Metadata & Steganography: Extract deep EXIF data and detect hidden scripts or high-entropy anomalies.
- 👤 Privacy Blur: Automatically detect and blur human subjects in reports using high-recall face detection.
- 📥 Reporting: Export all forensic findings as structured JSON or human-readable TXT files.
Ensure you have Python 3.10+ installed on your system.
git clone https://github.com/Mr-Infect/image-forensic-toolkit.git
cd image-forensic-toolkitIt is recommended to use a virtual environment:
python3 -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activatepip install -r requirements.txtBefore running the application, you must cache the necessary AI models. You will need a Hugging Face token.
python3 download_models.py(Progress will be displayed in the terminal during download)
Once the models are downloaded, launch the Streamlit application:
streamlit run app.pyOn the first run, navigate to the About section in the app and click "Re-index Database Images". This will process the images in the images/ folder for similarity search.
app.py: Main Streamlit application.modules/: Core forensic logic components.images/: Centralized folder for image database.download_models.py: Utility script for pre-caching models.
Developed for Image Forensic & Evidence Analysis.